Support of store brands on Open Prices

[raphael0202]

This discussion stems from this PR on Product Opener.

It would be interesting to know the store brand (ex: Carrefour, Monoprix,...) associated with a specific store. It would be tremendously helpful to answer questions such as :

• Is store chain A more expensive than store chain B for this product?
• I want to display all prices of chain A for this product
• Chain A only sell organic products, I want the "organic" checkbox to be toggled by default when adding a raw product

As CloCkWeRX started to do, we can import OSM name-suggestion-index database to support this features.
I would suggest creating a new store_brands table where we would import the DB (name of the store brand, countries, wikidata ID).

[CloCkWeRX]

Alrighty! Quick starting points.

Where are all of the stores?

https://alltheplaces.xyz/ is a CC-0 web scraped index of many, many things with a lat, lon and link to the NSI - for example, shop/supermarket, shop/convience. The majority of these assert a brand:wikidata attribute.

Different jurisdictions have different opinions on "can facts be copyrighted", and these are the claims from the stores or chains themselves - so are they suitable to merge into an ODBL source? How good is the data? Etc. It's nuanced!

Some have built comparison tools, to enable mapping where the jursidiction aligns - for example @matkoniecz has done some good work making tools for "is this X in OpenStreetMap? The website says it's phone number is Y, is that correct?"
https://matkoniecz.codeberg.page/improving_openstreetmap_using_alltheplaces_dataset/

Okay, what if I just want ODBL, high quality data?

(AI generated query samples here, but they are gonna be close)

Depending on where you pull your OSM supermarkets/etc from you may just be able to import brand:wikidata at the same time. Below assumes overpass.
Use an overpass turbo query to select all supermarkets, convience stores and also pull the brand:wikidata.

[out:json][timeout:25];
// Gather all supermarkets with a brand:wikidata tag
(
  node["shop"="supermarket"]["brand:wikidata"];
  way["shop"="supermarket"]["brand:wikidata"];
  relation["shop"="supermarket"]["brand:wikidata"];
);
out body;
>;
out skel qt;

Working sample: https://overpass-turbo.eu/s/27dH

You can then go further and pull the wikidata descriptions of brands or other details via their SPARQL endpoint.

SELECT ?supermarket ?supermarketLabel ?supermarketDescription ?website ?phone ?storeLocator WHERE {
  # Find things that are a subclass or instance of supermarket
  ?supermarket wdt:P31/wdt:P279* wd:Q213441;  # instance of (or subclass of) supermarket
               wdt:P17 wd:Q408.              # located in Australia

  OPTIONAL { ?supermarket wdt:P856 ?website. }           # official website
  OPTIONAL { ?supermarket wdt:P1329 ?phone. }            # phone number
  OPTIONAL { ?supermarket wdt:P973 ?storeLocator. }      # described at URL (store locator often lives here)

  SERVICE wikibase:label {
    bd:serviceParam wikibase:language "en".
  }
}
ORDER BY ?supermarketLabel

https://query.wikidata.org/#SELECT%20%3Fsupermarket%20%3FsupermarketLabel%20%3FsupermarketDescription%20%3Fwebsite%20%3Fphone%20%3FstoreLocator%20WHERE%20%7B%0A%20%20%3Fsupermarket%20wdt%3AP31%2Fwdt%3AP279%2a%20wd%3AQ213441%3B%20%20%23%20instance%20of%20%28or%20subclass%20of%29%20supermarket%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20wdt%3AP17%20wd%3AQ408.%20%20%20%20%20%20%20%20%20%20%20%20%20%20%23%20located%20in%20Australia%0A%0A%20%20OPTIONAL%20%7B%20%3Fsupermarket%20wdt%3AP856%20%3Fwebsite.%20%7D%20%20%20%20%20%20%20%20%20%20%20%23%20official%20website%0A%20%20OPTIONAL%20%7B%20%3Fsupermarket%20wdt%3AP1329%20%3Fphone.%20%7D%20%20%20%20%20%20%20%20%20%20%20%20%23%20phone%20number%0A%20%20OPTIONAL%20%7B%20%3Fsupermarket%20wdt%3AP973%20%3FstoreLocator.%20%7D%20%20%20%20%20%20%23%20described%20at%20URL%20%28store%20locator%20often%20lives%20here%29%0A%0A%20%20SERVICE%20wikibase%3Alabel%20%7B%0A%20%20%20%20bd%3AserviceParam%20wikibase%3Alanguage%20%22en%22.%0A%20%20%7D%0A%7D%0AORDER%20BY%20%3FsupermarketLabe

Note, this is too broad as it pulls non food businesses.
Note 2, you can call an API endpoint and get all this as structured data. You can query by individual wikidata ID as well.

There is even a (reasonably) maintained index of "store locator pages" in Wikipedia, so if users see product X is sold by chain Y, they can with a few clicks work out where to go.

[CloCkWeRX]

I will generate some code for this, hopefully over the weekend - is there a current spot where you lookup all OSM stores?

[matkoniecz]

Depending on where you pull your OSM supermarkets/etc from you may just be able to import brand:wikidata at the same time. Below assumes overpass.

Overpass is great for prototyping and small scale use, but for global processing getting planet data is going to work better (it is not as scary as it seems, and can be done even on low-end laptops - if unsure how I can dig out code I made for that comparison tool that got mentioned)